CN-122021447-A - Hydrological model calibration method based on remote sensing evapotranspiration space-time double constraints

CN122021447ACN 122021447 ACN122021447 ACN 122021447ACN-122021447-A

Abstract

The invention discloses a hydrological model calibrating method based on remote sensing evapotranspiration space-time double constraint, which comprises the steps of 1) collecting hydrological, meteorological and remote sensing data in a research area, 2) preprocessing the data, 3) constructing a spatial similarity index, incorporating a local Morgan index into spatial similarity evaluation, 4) constructing a hydrological model step-by-step calibrating framework, firstly, primarily calibrating a hydrological model by utilizing the spatial distribution of the remote sensing evapotranspiration data based on the spatial similarity index and optimizing a parameter interval, and secondly, calibrating the hydrological model by utilizing the time sequence of the remote sensing evapotranspiration data based on the optimized parameter interval to obtain a parameter final result. The invention carries out multi-step calibration on the hydrologic model based on the spatial distribution and the time sequence of the remote sensing evapotranspiration data, fully considers the space-time information contained in the remote sensing evapotranspiration data, reduces the uncertainty in the calibration process of the hydrologic model of the river basin, and can provide scientific basis for hydrologic simulation of the data-missing area.

Inventors

ZOU LEI

Assignees

中国科学院地理科学与资源研究所

Dates

Publication Date: 20260512
Application Date: 20260206

Claims (10)

1. The hydrological model calibration method based on remote sensing evapotranspiration space-time double constraint is characterized by comprising the following steps of: S1, collecting data, namely collecting hydrologic, meteorological and remote sensing data of a research area, wherein the remote sensing data comprise actual evapotranspiration data; s2, preprocessing data; s3, constructing a spatial similarity index, namely measuring the spatial autocorrelation of the evapotranspiration data by using a local Morlan index to obtain a local Morlan index matrix of the evapotranspiration data, and evaluating the similarity between the local Morlan index matrix simulating evapotranspiration and remote sensing evapotranspiration through cosine similarity, wherein the formula is as follows: Wherein c is cosine similarity between local Morgan index matrixes; And (3) with The local Morlan index value of the ith grid in the remote sensing evapotranspiration and simulated evapotranspiration spatial distribution matrix is respectively given, wherein n is the total number of the grids; the improved spatial similarity index SPSI is obtained by the following formula: Wherein a and b are respectively the ratio of the pearson correlation coefficient and the variation coefficient between the remote sensing evapotranspiration and the simulated evapotranspiration spatial distribution matrix, X i and Y i are respectively the evapotranspiration values on a grid i in the remote sensing evapotranspiration and the simulated evapotranspiration spatial distribution matrix, CV X and CV Y are respectively the variation coefficients of the remote sensing evapotranspiration and the simulated evapotranspiration spatial distribution matrix, mu X and sigma X are respectively the mean value and the variance of the remote sensing evapotranspiration spatial distribution matrix, and mu Y and sigma Y are respectively the mean value and the variance of the simulated evapotranspiration spatial distribution matrix; And S4, primarily calibrating a hydrologic model by utilizing the spatial distribution of the remote sensing evapotranspiration data based on SPSI indexes, optimizing a parameter interval, calibrating the hydrologic model by utilizing the time sequence of the remote sensing evapotranspiration data based on KGE coefficients by utilizing the optimized parameter interval, and obtaining a model parameter final result.
2. The method for calibrating the hydrological model based on the remote sensing evapotranspiration space-time double constraint of the invention is characterized in that in S1, the hydrological data comprise site actual measurement runoff data in a research area, the meteorological data comprise precipitation, air temperature, air pressure, sunshine hours and relative humidity data in the research area, and the remote sensing data further comprise leaf area indexes and land utilization/coverage types.
3. The hydrological model calibration method based on remote sensing evapotranspiration space-time double constraint of claim 1 is characterized in that in S2, the data preprocessing comprises spatial interpolation and resampling, the spatial interpolation adopts an inverse distance weighted average method for processing non-valued areas in meteorological and remote sensing data, the resampling adopts a bilinear interpolation method for uniformly resampling collected meteorological and remote sensing data to 0.05-degree resolution to uniformly obtain the resolution of the data.
4. The method for calibrating a hydrological model based on remote sensing evapotranspiration space-time double constraint of claim 1, wherein in S3, the calculation of the local Morgan index is as follows: wherein I i is the local Morganella index value of the ith grid in the evapotranspiration spatial distribution matrix, x i and x j are the evapotranspiration values on grid I and grid j respectively, and omega i,j is the spatial weight value between grid I and grid j; Is the mean value of the evapotranspiration space distribution matrix; average of the sum of squares of the differences between all grid evaporative values and their average.
5. The method for calibrating a hydrological model based on remote sensing evapotranspiration space-time double constraint of claim 4, wherein in S3, the space weight value between the grid i and the grid j is determined by an inverse distance weighting method: Where d ij is the Euclidean distance between grid i and grid j, and n is the total number of grids.
6. The method for calibrating the hydrological model based on the remote sensing evapotranspiration space-time double constraint of claim 1 is characterized in that in S4, the hydrological model is calibrated primarily by utilizing the space distribution of remote sensing evapotranspiration data, namely, the meteorological data, leaf area indexes in the remote sensing data and land utilization/coverage type data are input into the hydrological model to carry out hydrological simulation, the evapotranspiration space distribution in a research area is obtained through simulation, an optimization objective function is determined according to the space distribution of the remote sensing evapotranspiration and the simulated evapotranspiration in a simulation period, model parameters of a basin hydrological model are identified by utilizing an optimization algorithm, and initial model parameter values are obtained, wherein the optimization objective function is as follows: wherein SPSI ET is SPSI value between simulated evapotranspiration and remote sensing evapotranspiration spatial distribution, and the optimization algorithm comprises a genetic algorithm, a simplex method and a shuffling complex evolution algorithm.
7. The method for calibrating the hydrological model based on the remote sensing evapotranspiration space-time double constraint of claim 1, wherein in S4, the optimization parameter interval is performed by adopting the following formula: wherein: And (3) with Respectively taking a lower limit and an upper limit of the optimized value of the parameter k; And (3) with The lower limit and the upper limit of the original value of the parameter k are respectively adopted; The method is used for obtaining the initial value of the parameter k based on remote sensing evapotranspiration data space constraint.
8. The method for calibrating the hydrological model based on the remote sensing evapotranspiration space-time double constraint of claim 1 is characterized in that S4, the hydrological model is calibrated by utilizing a time sequence of remote sensing evapotranspiration data based on KGE coefficients, the hydrological model is subjected to hydrological simulation by inputting a leaf area index and land utilization/coverage type data in the remote sensing data into the hydrological model to obtain an evapotranspiration time sequence, an optimization objective function is determined according to a remote sensing evapotranspiration sequence value and a simulated evapotranspiration sequence value in a simulation period, model parameters of a watershed hydrological model are identified by utilizing an optimization algorithm, and final values of the model parameters are obtained, wherein the optimization objective function is as follows: Wherein KGE ET is KGE coefficient between the simulated evaporation and remote sensing evaporation sequences; The optimization algorithm comprises a genetic algorithm, a simplex method and a shuffling complex evolution algorithm; Wherein alpha, beta and gamma are respectively pearson correlation coefficient, mean ratio and variation coefficient ratio between the simulated evapotranspiration time sequence and the remote sensing evapotranspiration time sequence, O t and S t are respectively evapotranspiration values at the t-th moment of the remote sensing evapotranspiration and the simulated evapotranspiration time sequence, CV S and CV O are respectively variation coefficients of the simulated evapotranspiration and the remote sensing evapotranspiration time sequence, mu O and sigma O are respectively mean value and variance of the remote sensing evapotranspiration time sequence, mu S and sigma S are respectively mean value and variance of the simulated evapotranspiration time sequence, and N is a time period length.
9. The hydrological model calibration method based on remote sensing evapotranspiration space-time double constraint of claim 1 is characterized by further comprising the step of performing hydrological simulation based on obtained model parameters and model driving data, and evaluating a runoff simulation effect of a model, wherein the model driving data comprise meteorological data, leaf area indexes in the remote sensing data and land utilization/coverage type data.
10. The method for calibrating a hydrological model based on remote sensing evapotranspiration space-time double constraint of claim 9, wherein the runoff simulation result evaluation index comprises a KGE coefficient, a deterministic coefficient, a root mean square error and a percentage deviation.

Description

Hydrological model calibration method based on remote sensing evapotranspiration space-time double constraints Technical Field The invention relates to the technical field of watershed water circulation simulation, in particular to a hydrological model calibration method based on remote sensing evapotranspiration space-time double constraints. Background Water is a fundamental natural resource supporting life, ecosystem and human society, and research on watershed water circulation is critical to sustainable development of society. The watershed hydrologic model is used as a core tool of hydrologic simulation, is an important way for researching watershed hydrologic cycle and hydrologic process rules, and is an effective tool for solving the actual problems of water resource planning management, flood forecast, quantitative evaluation of climate change influence and the like. Due to the complex diversity of natural elements and hydrologic feature attributes within the watershed, hydrologic models often contain a large number of model parameters that are difficult to determine by direct measurement, and the parametric rating of the model becomes an essential step in hydrologic modeling and improving hydrologic simulation accuracy. Because runoff is easier to measure than other hydrologic elements and has higher accuracy, calibrating the hydrologic model by using measured runoff data is the most conventional means at present. However, this method is difficult to be applied in data-missing and data-free areas. The current hydrologic modeling for areas without data and without data mainly comprises two methods, namely a regional method for carrying out parameter transplanting based on hydrologic information of a river basin with a measuring station and a regional method for calibrating a hydrologic model based on multi-source remote sensing data. The regional method based on parameter transplantation needs to compare hydrologic characteristic factors of the measuring station watershed and the non-data watershed, relies on similarity between the watersheds, and has larger influence on the density of the measuring station. The remote sensing data has the advantages of wide coverage, strong space-time continuity and the like, and is widely applied to hydrologic simulation research. In recent years, various remote sensing products are continuously emerging and along with the continuous improvement of precision, and a calibration method based on remote sensing data is gradually becoming the main stream of hydrologic modeling in a non-data area. Evaporation is an important variable linking water circulation, carbon circulation and energy circulation. The remote sensing evapotranspiration data is utilized to rate the hydrologic model, so that the method has important reference value for runoff simulation of the lack-of-detection area and is widely applied. In the existing remote sensing evapotranspiration data calibration method, a model is used for parameter calibration by utilizing a time sequence of remote sensing evapotranspiration data. However, this method relies on the quality of the telemetry data, has some uncertainty, and ignores the spatial information characteristics of the telemetry data. Meanwhile, the model parameter calibration does not consider the time constraint and the space constraint of remote sensing evaporation to carry out step-by-step calibration, the space-time information contained in the remote sensing evaporation data cannot be fully utilized, the trade-off between different constraints is avoided, and the watershed water circulation simulation precision is further influenced. Disclosure of Invention The invention aims to provide a hydrological model calibration method based on remote sensing evapotranspiration space-time double constraint, which aims to provide scientific reference for watershed hydrologic cycle modeling of areas without data and without data. In order to achieve the above purpose, the present invention provides the following technical solutions: A hydrological model calibration method based on remote sensing evapotranspiration space-time double constraints comprises the following steps: s1, collecting data, namely collecting hydrologic, meteorological and remote sensing data in a research area, wherein the remote sensing data comprise actual evapotranspiration data; s2, preprocessing data; s3, constructing a spatial similarity index, namely taking the local Morand index into the evaluation of the spatial similarity, and providing the spatial similarity index taking the element spatial autocorrelation into consideration, wherein the spatial similarity index specifically comprises the following steps: 3-1 calculating a local Mortiered index of the evapotranspiration data by measuring spatial autocorrelation of the evapotranspiration data with the local Mortiered index and calculating a local Mortiered index matrix of the evapotranspiration data; the partial molan index is